Implement <linux/seqlock.h>.

(riastradh)

Declare struct seq_file; save some debugfs/sysfs ifdef trouble.

(riastradh)

Make <linux/mutex.h> side-load <linux/spinlock.h>.

Upstream actually side-loads <linux/spinlock_types.h> so you just get
spinlock_t without all the other spinlock operations, but whatever.

Reduce some diffs that this renders unnecessary.

(riastradh)

drm_waitqueue_t needs <drm/drm_wait_netbsd>.

(riastradh)

Make sure uapi drm.h defines __u* and __s* types even in kernel.

(riastradh)

Reduce some unnecessary diff, unused now.

(riastradh)

Implement kref_read.

(riastradh)

Move struct rb_root from linux interval_tree.h to rbtree.h.

(riastradh)

Define BUILD_BUG_ON_INVALID.

(riastradh)

struct timer_list needs <linux/timer.h>.

(riastradh)

drm_waitqueue_t needs <drm/drm_wait_netbsd>.

(riastradh)

ERR_PTR needs <linux/err.h>.

(riastradh)

ktime_t needs <linux/ktime.h>; need struct drm_unique forward decl.

(riastradh)

spinlock_t needs <linux/spinlock.h>; need forward decl of drm_crtc.

Disable waitqueue for now.  We won't be using the .c code; we just
need the .h file to compile.

(riastradh)

Move <linux/compiler.h> to common.

(riastradh)

nouveau_drm.h was renamed upstream to nouveau_drv.h.

(riastradh)

Delete nouveau_sysfs.c, removed upstream.

(riastradh)

Stub Linux <trace/events/dma_fence.h>.

(riastradh)

fence -> dma_fence in linux_reservation.c.

(riastradh)

Move IS_REACHABLE to <linux/kernel.h> to avoid double definition.

(riastradh)

Bunch more Linux header file stubs.

(riastradh)

No need for <linux/fence.h> in linux_dma_buf.c.

(riastradh)

Define IS_ENABLED and IS_REACHABLE on the command line.

Header file is too late for it, apparently.

XXX Find a way to put this into a header file; command-line -D
arguments are ugly.

(riastradh)

linux/sizes.h

(riastradh)

More Linux header stubs.

Let's just use empty header files until we actually populate these.
Save a bit of disk space...

(riastradh)

Define IS_ENABLED and IS_REACHABLE.

(riastradh)

<sys/stdarg.h> in NetBSD.

(riastradh)

Fix up #ifdefs.

(riastradh)

Stub linux interval_tree_generic.h, mem_encrypt.h, sync_file.h.

(riastradh)

fence -> dma_fence in <linux/reservation.h>.

(riastradh)

Stub <linux/srcu.h>.

(riastradh)

Include drm_internal.h by "../dist/drm/drm_internal.h".

(riastradh)

Stub linux compiler.h, scatterlist.h, seqlock.h, sort.h.

(riastradh)

No more hacks for dist/uapi -- got moved to dist/include/uapi.

(riastradh)

Stub <linux/irqreturn.h> to just include drm_irq_netbsd.h.

(riastradh)

Reimplement GPL <linux/llist.h> using only the upstream comments.

Hope I gathered what this is trying to do...!

(riastradh)

No more <drm/drm.h>.

(riastradh)

Rename fence -> dma_fence, step 3 of 3: code.

(riastradh)

Rename fence -> dma_fence, step 2: files files.

(riastradh)

Rename fence -> dma_fence, step 1: files.

(riastradh)

Stub i915_sw_fence.c.

(riastradh)

Disable GPL drm_writeback.c, minimal stub for drm_writeback.h.

(riastradh)

Disable *_trace_points.c.

I accidentally deleted these in the import, because the *_trace.h
files got in the way and I forgot the *_trace_points.c files were
fine.

There's nothing important in them anyway, so nothing of value is
lost.  Maybe next time I will remember to delete the *_trace.h files
on import but not the *_trace_points.c to reduce diffs, but that's
too much trouble at this point.

(riastradh)

Unfinished local implementation of GPL drm_lease.c.

(riastradh)

Local reimplementation of GPL drm_gem_framebuffer_helper.c.

(riastradh)

Allow maximum number of bits in DMA_BIT_MASK.

(riastradh)

Generate files.* using the 2netbsd scripts.

Author: Maya Rashish <maya@NetBSD.org>

(riastradh)

Handle more kinds of empty lines.

Author: Maya Rashish <maya@NetBSD.org>

(riastradh)

Break down and use gmake for i915drmkms2netbsd too.

Author: Maya Rashish <maya@NetBSD.org>

(riastradh)

Set CONFIG_DRM_VM=1, apparently needed for nouveau now.

Author: Maya Rashish <maya@NetBSD.org>

(riastradh)

Revert unnecessary typo fixes in upstream code.

This reverts most of the changes under sys/external/bsd/drm2/dist
from

https://mail-index.netbsd.org/source-changes/2020/12/16/msg125197.html

(Some of them were applied upstream or were in code that's gone now.)

(riastradh)

drm: Temporarily restore some local files deleted upstream.

These had local changes which needed to be distributed into other
files upstream, but that didn't happen in the merge commit.  It will
happen, and the files will be deleted, in a later commit.

(riastradh)

Merge linux-drm-v5-6-rc3

(riastradh)

new acpica

(christos)

merge changes from acpica-20210930 to acpica-20211217

(christos)

arm: Forward-declare `struct proc' before using in prototype.

(riastradh)

Add some missing includes to uvm_device.h.

- sys/types.h for dev_t
- sys/queue.h for LIST_ENTRY
- uvm/uvm_object.h for complete struct uvm_object type
- uvm/uvm_param.h for voff_t/vsize_t
- uvm/uvm_prot.h for vm_prot_t

(riastradh)

CTASSERT requires <lib/libkern/libkern.h>.

(riastradh)

arm: Need files.acpi for drm to discover NACPICA=0.

(riastradh)

Fix two bugs

i)  EHCI_HCSPARAMS should be read with EREAD4 (and not EOREAD4). Spotted by
    jmcneill

ii) Apply brackets so that ?: vs | operator precedence doesn't give the
    wrong result.

(skrll)

lint: document wrong data structures for modelling initializations

No functional change.

(rillig)

lint: move maximum seen subscript from brace level to initialization

There is no need to store this information at every brace level since in
any translation unit that survives a conforming C99 compiler, an array
of unknown size is only possible once per initialization, not once per
brace level.

(rillig)

lint: format grammar consistently

No binary change.

(rillig)

lint: use vector instead of list for designation in initialization

This change is a prerequisite for fixing the current bugs in handling
initializations.  Part of the fix will be designation_pop, which is
costly for a singly linked list.

As a side benefit, memory management becomes simpler and needs fewer
malloc calls.

No functional change.

(rillig)

Compensate for changes made in mixerctl.c rev 1.29
Usage msg now appears on stderr, and causes exit status to be 1

(kre)

only write the mask register if we're actually changing it

(macallan)

cg14_invert():
- use only 32bit accesses
- use byte mask / ROPs for partial writes

(macallan)

lint: reduce code for beginning an initialization

No functional change.

(rillig)

lint: inline sub-cases of brace_level_sub_type

Having the code in separate functions did not add to the clarity of the
code.  The additional information from the function names can be grasped
as easily from the case labels.

No functional change.

(rillig)

One build break fixed, another appears...

Use the correct pointer, not the one used in the other piece of
code that changed at the same time..

(kre)

tests/lint: add more examples for initialization, from C99

(rillig)

Remove an unnecessary test for NULL (the same thing is done in the
following lines) along with an incorrect return of nothing if it is
found (the following version correctly returns NULL).  Should unbreak build.

(kre)

fix error messages, usage, add static, more checks for ioctl, KNF

(christos)

Provide a clipping function and use it to prevent setting values out of bounds.
(from RVP)

(christos)

If the step is 0, use 1 instead of always incrementing step by one to avoid
zerodivide (from RVP)

(christos)

more protection from unset threads from RVP

(christos)

lint: fix initialization with few braces from function

Seen in Postfix, smtp_proto.c.

(rillig)

tests/lint: demonstrate another initialization with few braces

Seen in Postfix, smtp_proto.c.

(rillig)

lint: fix initialization of array with fewer braces

Seen in Postfix, cleanup_strerror.c.

(rillig)

tests/lint: demonstrate another initialization that lint cannot handle

Seen in external/ibm-public/postfix/dist/src/global/cleanup_strerror.c.

(rillig)

lint: work around lint bug in initialization with few braces

Needs a proper fix later, but for now, this workaround allows to run
lint on Postfix again.

(rillig)

Add missing WM_TX_{,INTR_}PROCESS_LIMIT_DEFAULT options same as RX side.

(knakahara)

add two additional cyperpower ups ids for ignore hid.

(mrg)

regen.

(mrg)

add two more models of cyperpower UPS.

(mrg)

In vattr_init_mask() check for va_flags and if set add to the mask of
to the things that setattr is interested in.
Fixes chflags(2) on ZFS.

Thanks riastradh@ for fix.

(simonb)

Correct copypaste comment grammar.

(skrll)

Trailing whitespace

(skrll)

curses: spell "foreground" correctly in comments

(uwe)

lint: rename is_string_array to is_character_array

The old name was too confusing.

No functional change.

(rillig)

lint: demonstrate wrong error message in initialization

History of the new test fragment:

libexec/2021.09.14.19.44.40-plain/lint1
| (14): error: cannot initialize 'struct typedef histogram_entry'
|    from 'pointer to char' [185]
| exit status 1
libexec/2021.04.01.14.20.30-plain/lint1

libexec/2021.03.30.20.23.30-plain/lint1
| (14): error: cannot initialize 'struct typedef histogram_entry'
|    from 'pointer to char' [185]
| (14): error: cannot initialize 'struct typedef histogram_entry'
|    from 'int' [185]
| (15): error: cannot initialize 'struct typedef histogram_entry'
|    from 'pointer to char' [185]
| (15): error: cannot initialize 'struct typedef histogram_entry'
|    from 'int' [185]
| exit status 1
libexec/2021.03.30.14.25.28-plain/lint1

libexec/2021.03.29.21.34.17-plain/lint1
| (15): error: too many struct/union initializers [172]
| exit status 1
libexec/2021.03.28.15.36.37-plain/lint1

libexec/2021.03.28.14.13.18-plain/lint1
| (18): error: too many struct/union initializers [172]
| exit status 1
libexec/2021.03.21.20.44.59-plain/lint1

libexec/2021.03.21.20.30.19-plain/lint1
| (18): too many struct/union initializers [172]
| exit status 1
libexec/2014.11.20.20.49.06-plain/lint1

libexec/2014.11.20.20.48.33-plain/lint1
| exit status 0

(rillig)

lint: in GCC mode, declare alloca and variants

The prototype declarations define the correct parameter types of these
functions so that they are no longer subject to the default argument
promotions (C11 6.5.2.2p6).

The GCC builtins are only recognized in GCC mode (-g).

(rillig)

lint: narrow down parameter of build_name

Passing an arbitrary tokenizer symbol left too much freedom and
uncertainty to the caller, and 0 was a magic number in this context.

No functional change.

(rillig)

Fix typo: reader -> header

(reinoud)

Fix serious issue with recordable media formatted with Win10. When closing it
after modification the VAT was written out corrupted making the disc
unreadable anymore on remount. Thanks for ig@ for spotting it in the wild!

(reinoud)

Make sysctl udf.verbose dependent on UDF_DEBUG instead of DEBUG

(reinoud)

tests/lint: ensure that tests on a message mention this message

(rillig)

tests/lint: fix test on platforms where char is unsigned

In tree.c 1.385 from 2021-10-09, the format of the message changed.  The
test for platforms where char is unsigned was not updated in that commit
because I had forgotten about it, and because lint1/check-msgs.lua only
checks the message in tests named exacty msg_000.c, but not msg_000_*.c.

Found by martin@.

(rillig)

Remove bogus architecture detection voodo - tsan is not available
on most architectures, so just state it that way.

(martin)

interlim commit to snapshot SC2A11 GbE progress.

(nisimura)

improve SC2A11 hardware register description.

(nisimura)

tests/lint: remove generated shell program on 'make clean'

(rillig)

Print NVM image version on 82598.

  FreeBSD ix-3.3.29 added code to decode NVM version. On 82598, the NVM
  offset is NVM_EEP_OFFSET_82598(== 0x2a). My own three different 82598
  cards' value in NVM_EEP_OFFSET_82598 are 0xffff. Instead, the dev starter
  version (0x29) has the value. Two of them have 0x1070 and another has 0x2090.
  The 82598 specification update notes about 2.9.0.

(msaitoh)

tests/lint: fix copy-and-paste typo in comment

(rillig)

lint: improve error message for missing newline in .ln file

(rillig)

tests/lint: demonstrate bad error message on missing newline

(rillig)

While searching the VAT on recordable media, search the last sector too!
This fixes Win10 formatted discs from being mounted incorrectly.

(reinoud)

Fix for PR lib/56298

Remove the DIAGASSERT for str being NULL in the puts/putp functions,
add protection so that the functions just return OK if str is NULL.
This prevents the assert firing when libcurses passes through a NULL
due to an undefined terminfo entry.

(blymn)

Grr: need NOLINT=yes for the Makefiles that include Makefile.xf86-driver
early, and MKLINT:=no for the ones that include it late. Should fix that.

(christos)

Lint2 is broken, so disable lint for now.

(christos)

Add -D__GNUC__ for all drivers.

(christos)

Add -D__GNUC__ for lint

(christos)

Add more attributes

(christos)

fix lint

(christos)

no lint (c++)

(christos)

-W === -Wextra

(christos)

make: amend leftover cleanups from the previous commits

No functional change.

(rillig)

make: format comments according to /usr/share/misc/style

Assisted by indent(1), with manual corrections due to its many remaining
bugs.

No functional change.

(rillig)

remove naked -W

(christos)

make: use consistent indentation for statements and continuations

No binary change, except for line numbers in assertions in suff.c.

(rillig)

make: remove redundant comments for multiple-inclusion guards

(rillig)

make: remove space after ':' in bit-field declarations

As seen in /usr/share/misc/style.

(rillig)

make: remove bmake_free

It was only used in 2 places, and in both these places, the pointer was
never null.

(rillig)

make: in CLEANUP mode, free interned strings at the very end

Noticed by sjg.

(rillig)

make: in ParseDependencyCheckSpec, move default branch to the bottom

No functional change.

(rillig)

make: change return type of HashTable_Set to void

None of the callers needs the HashEntry for further manipulation.

No functional change.

(rillig)

make: change return type of Compat_RunCommand from int to bool

The documentation was wrong before since status was not restricted to
only 0 or 1.

No functional change.

(rillig)

Change DCB credit parameters. No functional change.

  - From FreeBSD ix-3.3.24.
  - No functional change because NetBSD doesn't support DCB.

(msaitoh)

make: mark several functions whose result must be used

Suggested by sjg, to catch more bugs like the memory leak in cond.c
1.303 from 2021-12-13.

No binary change.

(rillig)

Remove dead code. No functional change.

  FreeBSD: 0b487fb4547ad1a939be0d523e555557c94b7cd1
  DPDK:    a6395d471e14e5a7432875dad8fb3533238c5167

(msaitoh)

Check host interface return status. No functional change on NetBSD.

  - FreeBSD: c1a56b6f5ffd6f3180a654d058c1783ccb808e8b or ix-3.3.18
    DPDK:    db18e37090a3b9af47d6a6886248520f6b220bf9

      Writing to read-only fields returns a non-OK Return Status
      for shadow RAM write command for X550.
      This information was previously discarded.

  - No functional change on NetBSD because our ixg has no interface to
    write the NVM.

(msaitoh)

make: prevent memory leaks from buffers

The warning about unused function results would have prevented the
memory leak that was fixed in cond.c 1.303 from 2021-12-13.

(rillig)

Match X550_PHY_ID correctly on X550.

  - Sync with FreeBSD ix-3.3.18.
    - phy_id's revision field is cleared, so use new X550_PHY_ID.
  - Before this commit, phy_id was set to ixgbe_phy_cu_unknown on X550.
    Now it's set to ixgbe_phy_aq.

(msaitoh)

When running the tests under "qemu -accel kvm" on a Linux host,
isQEMU_TCG() should return false.  Fixes multiple test cases that were
failing with "Test case was expecting a failure but none were raised"
on that platform.

(gson)

Improve PHY's dmesg output for non-MIIVERBOSE case.

(msaitoh)

Cosmetic change for the output of mii_get_descr().

(msaitoh)

Fix typo in comment.

(knakahara)

lint: unexport struct_or_union_member

Strangely, lint2 didn't notice this unnecessary export.

(rillig)

Disable lint, some of the code is c++ and some we can't handle __extension__
in struct in libctf ctf-error.c

(christos)

Don't install a symlink for when there's already a file.

Caused some cryptic issues in the builds far further along.
Pointed out by paulg, thanks.

(maya)

deal with alloca for lint.

(christos)

avoid __extension__ redefinition

(christos)

Another instance of transparent struct/union

(christos)

Avoid NULL deref if there is no current thread. From RVP.
Prints "No applicable message"

(christos)

Improve restrict handling.

(christos)

don't gut restrict for lint.

(christos)

Ticket #1393

(martin)

Pull up following revision(s) (requested by riastradh in ticket #1393):

sys/dev/cgd.c: revision 1.141

cgd(4): Wait for worker threads to complete before destroying mutex.

Fixes PR kern/56546 (probably!).

(martin)

tests/lint: demonstrate missing support for anonymous members

(rillig)

provide an expanded description for -W

(christos)

Disable lint and explain why.

(christos)

Add expected-fail test for kern/56549

consecutive sem_open() calls do not return the same address

(wiz)

Also accept -W to pass to cpp

(christos)

Pass W with IDU for lint and deps.

(christos)

-Wno-comment for everyone (including lint)

(christos)

PR/56548: RVP: mail(1) segfaults if thread commands are used in ~/.mailrc

(christos)

lint fixes:
1. avoid typedef re-definitions (perhaps introduce c11 mode for lint)
2. add __lint__ wherever __GNUC__ is defined.

(christos)

don't conditionalize the installation of nouveau directories

(maya)

ffs: fix the creation of device nodes on file systems with ACLs enabled.

(chs)

ffs: support extattrs (and thus ACLs) on fifos.

(chs)

Use adapter->mta for the multicast array memory instead of the on-stack array.

  - Same as ixv-1.5.22.
  - Same as ixg(4).

(msaitoh)

Wait longer for link after fiber MAC setup.

  - FreeBSD:
        aaa3af802f90d93fdffb99100fe56f0bc3dda119
        96ef6eb3ae9d622906fb838c82ede3074f864cdc

    DPDK:
        64f1c8539c8ce99214b9eb1fb728a2c6745f3300
        73247f1ced303c16987bb366d38a2d8a0fc40db4

    After setting up the link on a fiber port, the maximum wait time for
    the link to come up is 500 ms in ixgbe_setup_mac_link_multispeed_fiber().
    On an x550 SFP+ port, this is often not sufficiently long for the link
    to come up. This can result in never being able to retrieve accurate
    link status for the port using rte_eth_link_get_nowait().

    Increase the maximum wait time in ixgbe_setup_mac_link_multispeed_fiber()
    to 1 s.

(msaitoh)